To improve communication efficiency during the updating of a replica, and to reduce the amount of memory required for the updating of a replica. The present invention is a computer for maintaining consistency of replica contents by interchanging data modification with another computer. The computer includes: a replica including shared data; a receiver for receiving, from another computer, a modification request for data included in the replica; and a controller for controlling, regardless of a receiving order of the modification request, for each receipt of the modification request, a timing of update execution for the replica in accordance with the received modification request by using information included in the modification request. The timing of replica updating in accordance with the received modification requests can be determined, regardless of the order in which the modification requests are received, that is, the order of the transmission from other computers. Therefore, communication efficiency can be improved and the amount of required memory can be reduced.
|
11. A method for maintaining consistency replica contents between computers, said method comprising:
a receiver in a first computer for receiving a modification request from a second computer for data included in a replica having shared data, wherein said modification request includes: a designation indicating no restriction for a modification order; a first designation indicating that after all modification requests preceding a current modification request have been affected by said replica, said replica is to be updated in accordance with said current modification request; a second designation indicating that modification requests following a current modification request are to be affected by said replica after said replica has been updated in accordance with said current modification request; or a third designation indicating that after all modification requests preceding a current modification request have been affected by said replica, said replica is to be updated in accordance with said current modification request, and that all modification requests following said current modification request are to be affected by said replica after said replica has been updated in accordance with said current modification request; and controlling regardless of a receiving order of said modification request, for each receipt of said modification request, a timing of update execution for said replica in accordance with the received modification request by using information included in said modification request. 1. A computer for maintaining consistency of replica contents by interchanging data modification with another computer, said computer comprising:
a replica including shared data; a receiver for receiving, from said another computer, a modification request for data included in said replica, wherein said modification request includes: a designation indicating no restriction for a modification order; a first designation indicating that after all modification requests preceding a current modification request have been affected by said replica, said replica is to be updated in accordance with said current modification request; a second designation indicating that modification requests following a current modification request are to be affected by said replica after said replica has been updated in accordance with said current modification request; or a third designation indicating that after all modification requests preceding a current modification request have been affected by said replica, said replica is to be updated in accordance with said current modification request, and that all modification requests following said current modification request are to be affected by said replica after said replica has been updated in accordance with said current modification request; and a controller for controlling, regardless of a receiving order of said modification request, for each receipt of said modification request, a timing of update execution for said replica in accordance with the received modification request by using information included in said modification request. 6. A computer for maintaining consistency of replica contents by interchanging modification data with another computer, said computer comprising:
a replica including shared data; means for instructing modification of data included in said replica, and for designating information concerning a modification order to the modification of data, wherein said information concerning said modification order includes: a designation indicating no restriction for a modification order; a first designation indicating that after all modification requests preceding a current modification request have been affected by said replica, said replica is to be updated in accordance with said current modification request; a second designation indicating that modification requests following a current modification request are to be affected by said replica after said replica has been updated in accordance with said current modification request; or a third designation indicating that after all modification requests preceding a current modification request have been affected by said replica, said replica is to be updated in accordance with said current modification request, and that all modification requests following said current modification request are to be affected by said replica after said replica has been updated in accordance with said current modification request; means for generating a modification request that corresponds to an instruction of the modification of data and includes information concerning said modification order; and a transmitter for transmitting said modification request in such a manner that it is not confirmed that a generation order and a transmitting order of said modification request are identical.
2. The computer according to
a first number of modification requests that have been generated and that include said second or third designation; and a second number of modification requests that have been generated following the last modification request including said second or third designation.
3. The computer according to
4. The computer according to
5. The computer according to
7. The computer according to
a first number of modification requests that have been generated and that include said second or third designation; and a second number of modification requests that have been generated following the last modification request including said second or third designation.
8. The computer according to
9. The computer according to
10. The computer according to
12. The method according to
a first number of modification requests that have been generated and that include said second or third designation; and a second number of modification requests that have been generated following the last modification request including said second or third designation.
13. The method according to
14. The method according to
15. The method according to
|
The present invention relates to a system wherein computers have replicas of shared data and maintain replica content consistency by interchanging data modifications, more particularly to the order control of the replica updating. A conventional data sharing system uses the following method to determine the order of the updating of a replica.
In many cases, a data sharing system, in which it is assumed that computers are constantly connected by communication lines, includes a mechanism for issuing a notification each time data are updated. In this case, it is ensured that each time data are modified the order of replica updating will correspond to the sequential order of the data modifications. Such a sequential process is hereinafter referred to as the FIFO order. However, there are cases where portable terminals at remote, outside of the office, are used to read out, browse and/or edit data in computers installed in offices; and as the communication quality provided by the wireless public networks that are used for such portable terminals is lower than is that provided by normal LANs (Local Area Networks), there are frequent packet transmission failures. Then, generally, as the restrictions imposed on the data transferring order become stricter, this is accompanied by a corresponding increase in the costs for a transfer. The FIFO order transmission with the strictest restrictions takes the large overhead. Since using a wireless public network is in itself expensive, this transfer overhead constitutes a serious problem.
In a data sharing system, such as Lotus Notes (a trademark of Lotus Development Corp.), for which it is assumed that operations are also performed during the disconnection of the communication lines, a one-time processing, similar to a batch processing, is used, to update the replica. With this system, data can not be accessed during updating. In addition, when the updating of the replica is interrupted by the disconnection of the communication line, the order in which the updating process was performed is generally indefinite. That is, it is unknown to which data has been updated. EP0794646 (Japanese Published Unexamined Patent Application No. Hei 9-24642) discloses a data management system for managing copies of shared data file, which are maintained in a plurality of computer systems connectable via a mobile communication network. The data management system comprises: logging means for storing records of modifications that are correlated with each copy of the shared data file and that are executed for these copies; extraction means for extracting, via a connection to the mobile communication network, the records stored for the other copies of the shared data file; merging means for merging the extracted records to generate the order of the modifications; conflict resolving means for resolving conflicts in the order of the modification by applying a rule defined in advance to the modification order; and means for modifying the copies of the shared file in accordance with the modification order in which conflicts are resolved. In this case, it is considered that the conflict means there are a plurality of data modifications which are performed in parallel for data in the same shared file edition. In addition, it is considered that the resolution of the conflict means that a plurality of data modifications in conflict are rearranged in a specific order. The data management system is provided to merely resolve the modification conflicts. In order to resolve the conflict, data concerning all the modifications must be prepared by the time a determination as to whether a conflict has occurred. That is, a copy of the shared data file can not be modified unless the modification conflict has been resolved. Therefore, if even one set of data concerning a modification is missing, the modification process can not be performed. In addition, the load imposed on the memory capacity is increased to store the modification related data that is received. This constitutes an immense problem for a portable terminal that has only a small memory capacity available for communication. Further, another problem arises if data must be re-transmitted because of a memory capacity shortage.
As described above, conventionally, during the updating of a replica problems arise concerning communication efficiency, the amount of required memory, and the process when the disconnection of a communication line happens during the data transmission.
It is, therefore, one object of the present invention to provide a method and an apparatus for improving communication efficiency during the updating of a replica.
To achieve the above objects, the present invention is a computer for maintaining consistency of replica contents by interchanging data modification with another computer. The computer includes: a replica including shared data; a receiver for receiving, from another computer, a modification request for data included in the replica; and a controller for controlling, regardless of a receiving order of the modification request, for each receipt of the modification request, a timing of update execution for the replica in accordance with the received modification request by using information included in the modification request. The timing of replica updating in accordance with the received modification requests can be determined, regardless of the order in which the modification requests are received, that is, the order of the transmission from other computers. Therefore, communication efficiency can be improved and the amount of required memory can be reduced.
The information included in the modification requests includes: a designation indicating no restriction for a modification order (Ordinary in the preferred embodiment);
a first designation indicating that after all modification requests preceding a current modification request have been effected for the replica, the replica is to be updated in accordance with the current modification request (Forward Flush in the embodiment); a second designation indicating that modification requests following a current modification request are to be effected for the replica after the replica has been updated in accordance with the current modification request (Backward Flush in the embodiment); or a third designation indicating that after all the modification requests preceding a current modification request have been effected for the replica, the replica is to be updated in accordance with the current modification request, and that all modification requests following the current modification request are to be effected for the replica after the replica has been updated in accordance with the current modification request (Two Way Flush). In this manner, the order of replica updating can be explicitly designated.
The information included in the modification request may further include: the number (a first number: BT value in the preferred embodiment) of modification requests that have been generated and that include the second or third designation; and the number (a second number: SBT value in the preferred embodiment) of modification requests that have been generated following the last modification request including the second or third designation. Then, a mechanism can be provided that, even if a communication line has been disconnected before the completion of the updating, can confirm to which modification requests have been received and used to update the replica.
The controller may store first and second management values (the values of BT and SBT in a process in the preferred embodiment) that respectively correspond to the first and second numbers and that define a timing of update execution for the replica. Therefore, the timing of the replica updating can be determined more easily.
If the replica is updated in accordance with the modification request, the controller may store, as third and fourth numbers (the values of response BT and SBT), the first and second numbers included in the modification request. The computer may further comprise a transmitter for transmitting to another computer a modification request for data in the replica and the controller may make the third and fourth numbers to be included in the modification request. As a result, even if a communication line is disconnected, the transmission source can determine which modification request is to be re-transmitted, and also which modification request is to be discarded.
The configuration of a computer on the receiving side has been explained. A computer on the transmission side includes: a replica including shared data; means for instructing modification of data included in the replica, and for designating information concerning a modification order to the modification of data; means for generating a modification request which corresponds to an instruction of the modification of data and which includes information concerning the modification order; and a transmitter for transmitting the modification request in such a manner that it is not confirmed that a generation order and a transmitting order of the modification request are identical. As described above, information concerning the modification order is clearly designated for each data modification, and is included in a modification request so that the receiving side can use that information. In addition, in order to improve communication efficiency, this information is transmitted in a manner that it is not confirmed that the generation order and the transmitting order of the modification request are identical. This manner includes a case where a communication line is disconnected and the re-connection of the line and the re-transmission of information are performed.
The modification request can include the same information as included in the previously described modification request.
The computer further may comprise a receiver for receiving from another computer a modification request for data included in the replica and the received modification request may include response information (the values of response BT and response SBT in the preferred embodiment) indicating that the replica has been updated by a destination computer in accordance with the transmitted modification request. Thus, a modification request is ensured that does not need to be re-transmitted. That is, the computer further comprises means for referring to the response information to determine whether the modification request that has been transmitted is to be deleted.
A single computer can include the elements of both the transmission side and the receiving side, and such computers may be used to constitute a data sharing system.
The configuration of a computer has been explained, and the processing to be performed by the computer can be provided as a program. This program can be stored on a floppy disk or a CD-ROM, or in another storage device.
An application 5 issues a request to the replica controller 7 for the generation, the searching, the referencing and the deletion of data and etc. When issuing a data modification request, such as the generation or the deletion of data, the order of the replica in the replica updating is designated. According to the present invention, the designations used for the order of replica updating are Ordinary, Forward Flush, Backward Flush and Two Way Flush. Since the access to data, other than the modification, is performed the same as in a normal database access, no detailed explanation for it will be given.
The Ordinary designation is used to indicate that no restriction is placed on the order of a corresponding replica updating when the modification is performed. The Forward Flush designation is used to indicate that the replica updating corresponding to the modification of the data in the replica will be performed after all preceding modifications have been effected for the replica. The Backward Flush designation is used to indicate that after the current data modification has been effected for the replica, succeeding modifications of data are to be effected for the replica. The Two Way Flush designation is used to indicate that the restrictions provided for both the Forward Flush and the Backward Flush designations apply. The four designations given above are described as to the order of transmission/reception message, for example, in "An Implementation of F-channels," Mohan Ahuja, IEEE Transactions on Parallel and Distributed Systems, Vol. 4, No. 6, June 1993. However, the application of the designations for a data sharing system is neither described nor taught in the referenced paper. Further, the portions related to a transmission queue and a reception queue, which will be described later, are also neither described nor taught in the paper.
Upon receiving an access request from the application 5, the replica controller 7 accesses to data stored in the data storage device 17, and adds a packet transmission request to the transmission queue 9. In addition, the replica controller 7 examines the state of the reception queue 11, and in accordance with a designated timing, updates the replica so as to correspond to the modification request stored in the reception queue 11. The replica controller 7 can be divided into a section for accessing to the data storage device 17 and a section for designating the contents to be accessed. According to the present invention, a conflict caused by the applications 5, and a conflict that occurs between a modification of data by an application 5 and a data modification request stored in the reception queue 11 are not handled, and the modification of data by the application 5 and the execution of the data modification request stored in the reception queue 11 are performed independently. If a conflict occurs, it may be resolved by using a conventional method, or either of the executions may be ignored.
The data storage device 17 is used to store a replica, and has the same functions as an ordinary database. That is, in accordance with a request received from the replica controller 7, data are generated, searched for, referred to, or deleted.
The transmission queue 9 is used to manage a packet transmission request. The packet transmission request includes the type and the contents of the data modification prepared by the application 5, the designation of the updating order, and the value of BT and the value of SBT when a transmission request is generated. The value of BT and the value of SBT are included in the transmission queue 9 as data for the management of the queue. The BT value is used to represent the number of data modifications (also called data modification requests) that have been received from the application 5 and for which either the Backward Flush or the Two Way Flush has been designated as the updating order. The BT value is managed by the replica controller 7. The SBT value is used to represent the number of data modifications generated from the time either the last Backward Flush or the last Two Way Flush was designated until the latest modification was generated. The value of SBT is also managed by the replica controller 7.
Other information given includes the value of response confirmation BT for managing the BT value for the latest modification request that is received by the computer B(3) and whose execution is performed for the replica; and the value of response confirmation SBT for managing the SBT for the latest modification request that is also received by the computer B(3) and whose execution is performed for the replica. These data are managed by the packet receiver 15.
The reception queue 11 is used to manage the modification requests in packets with which updating has not yet been effected, even though the packets have been received. In addition to the modification requests, there are the value of BT in process, which represents the value of the modification request BT that can currently be executed; the value of SBT in process, which represents the value of the modification request SBT that can be currently executed; the value of response BT, which represents the value of BT included in the latest modification request processed by the replica controller 7; and the value of response SBT, which represents the value of SBT included in the latest modification request processed by the replica controller 7. These numbers are managed by the replica controller 7.
The packet transmitter 13 monitors the transmission queue 9 and the reception queue 11, generates a necessary packet, and transmits it to a communication medium 19. The packet transmitter 13 sequentially transmits the transmission requests, stored in the transmission queue 9, in such a manner that it is not confirmed that the computer B(3) will receive the requests in the order in which they were stored in the transmission queue 9 (the generation order). Specifically, transmission is not performed in such a manner that a packet is transmitted after it has been confirmed that the preceding packet has been received in the generation order, but packets are sequentially transmitted. When the transmission destination computer B(3) fails to receive a packet, the pertinent packet is re-transmitted. The same packet may be repeatedly transmitted until it is confirmed that the packet has been received and the updating process has been performed. At this time, the value of response confirmation BT and the value of response confirmation SBT can be used. The transmitting order may not be at all related to the generation order of the transmission requests.
The packet transmitter 13 selects one of the packet transmission requests in the transmission queue 9, and prepares and transmits a packet which includes the value of response BT and the value of response SBT in the reception queue 11. The value of response BT and the value of response SBT are used to change the values of the response confirmation BT and SBT that are stored in the transmission queue 9 of the computer B(3). If no transmission requests are stored in the transmission queue 9, only the values of the response BT and SBT are transmitted.
The packet receiver 15 examines the contents of a received packet and the states of the transmission queue 9 and of the reception queue 11, and updates these queues. That is, the packet receiver 15 uses the value of response BT and the value of response SBT in a received packet to update the values of the response confirmation BT and SBT stored in the transmission queue 9. If the response BT>the response confirmation BT, the values of the response confirmation BT and SBT are updated by using the values of the response BT and SBT. If the response BT=the response confirmation BT, it is checked whether the response SBT>the response confirmation SBT. If this condition is established, the value of the response confirmation SBT is updated by using the value of the response SBT. If the above condition is not established, the values of the response confirmation BT and SBT are not updated.
The modification requests in the received packets may be duplicated. Therefore, the packet receiver 15 examines the value of BT and the value of SBT included in a modification request in a received packet. Then, if a modification request stored in the reception queue 9 has the same value of BT and the same value of SBT, the modification request in the packet received later is discarded.
Also, the modification request included in a received packet may be a request that has already been processed. Thus, the packet receiver 15 will discard, by referring to the reception queue 11, a modification request for which the value of included BT is smaller than that of the response BT, and a modification request for which the value of included BT is the same as that of the response BT and the value of included SBT is smaller than that of the response SBT.
A transmission request for which it is established that the response confirmation BT≧the included BT, or that the response confirmation BT=the included BT and the response confirmation SBT≧the included SBT, can be deleted from the transmission queue 9. The transmission request may be deleted by either the packet receiver 15 or the transmitter 13.
The BT and the SBT included in a modification request concerning a received packet are examined, and the modification requests stored in the reception queue 11 are rearranged in the ascending order of the BT and the SBT. This process may also be performed by either the packet receiver 15 or the replica controller 7.
The communication medium 19 may be either wireless or cable, or may be a combination of the two. The feature of the present invention lies in its ability to cope with problems that arise when data is transmitted via a less reliable communication network.
Referring to
Referring to
When at (4) Two Way Flush is designated, (1, 0) is assigned for data modification T1. In this case also, the value of the SBT is cleared and the value of the BT is incremented by one. When at (5) Ordinary is designated, (2, 0) is assigned for data modification O3, and then the value of the SBT is incremented by one. When at (6) Ordinary is also designated, (2, 1) is assigned for data modification O4, and the value of the SBT is incremented by one. When at (7) Forward Flush is designated, (2, 2) is assigned for data modification F1, and the value of the SBT is incremented by one as defined.
An explanation will now be given, referring to
First, the value of BT and the value of SBT in process are initialized to 0. At (1) the packet receiver 15B receives data modification O1 and adds it to the reception queue 11B. The state of the reception queue 11B is shown at the left in
(a) a case where Ordinary or Backward Flush is designated as the updating order of data modification requests
The pertinent data modification can be executed if:
the value of BT included in the data modification request≦the value of BT in process.
(b) a case where Forward Flush or Two Way Flush is designated as the updating order of data modification requests
The pertinent data modification can be executed if:
the value of BT included in the data modification request=the value of BT in process; and the value of SBT included in the data modification request=the value of SBT in process.
At (1) in
Then, it is checked whether the data modification O1 can be deleted from the reception queue 11B. This determination is performed by whether the value of BT included in the data modification request=the value of BT in process and the value of SBT included in the data modification request=the value of SBT in process. If the above condition is established, the values of BT and SBT in process are also updated. The updating method is performed as follows:
(a) a case where Ordinary or Forward Flush is designated as the updating order of data modification requests
the value of SBT in process=the value of SBT in process+1.
(b) a case where Backward Flush or Two Way Flush is designated as the updating order of data modification requests
the value of BT in process=the value of BT in process+1; and
the value of SBT in process=0.
Since the above determination reference is satisfied, the replica controller 7B can delete the data modification O1 from the reception queue 11B. Therefore, in
Following this, at (2) the packet receiver 15B receives data modification B1. Since Backward Flush is designated, the value of BT included in the data modification B1 is compared with the value of BT in process to determine whether the determination reference is satisfied. Thus, the replica controller 7B executes the data modification B1. "o" is entered in the "modification execution" row at (2) in FIG. 4. However, since the value of SBT in the data modification B1 differs from the value of SBT in process, the data modification B1 can not be deleted from the reception queue 11B. As shown at (2) in
At (3) the packet receiver 15B receives data modification O2. The value of BT included in the data modification O2 is 0 and the value of SBT is 1. Since the value of SBT is smaller than that in the data modification B1, as shown at (3) in
Since Ordinary is designated as the updating order and since the two values of the BT and the BT in process match and the two values of the SBT and the SBT in process also match, the value of SBT in process is incremented by one. In this case, the value of the BT in process is 0 and the value of the SBT in process is 2. Therefore, the values of the BT and the SBT included in the data modification B1 that is maintained in the reception queue 11B match the values of the BT in process and the SBT in process. The replica controller 7B then deletes the data modification B1 in the reception queue 11B. The deletion of the data modification B, is indicated by an arrow pointing from (3) to (2) in FIG. 4. Further, the right side at (3) in
At (4) the packet receiver 15B receives a data modification T1. Since the value of the BT included in the data modification T1 matches the value of the BT in process, and since the value of the SBT in the data modification T1 matches the value of the SBT in process, the replica controller 7B executes the data modification T1. "o" is entered in the "modification execution" row at (4) in FIG. 4. Further, since the data modification T1 can be deleted, the replica controller 7B deletes it from the reception queue 11B. "o" is entered in the "deletion" row at (4) in FIG. 4. Also at (4) in
At (5), the packet receiver 15B receives a data modification O4. Since Ordinary is designated as the updating order and both the value of the BT in the data modification O4 and the value of the BT in process are 2, the replica controller 7B executes the data modification O4. "o" is entered in the "modification execution" row at (5) in FIG. 4. However, since the value of the SBT in process does not match the value of the SBT included in the data modification O4, the data modification O4 can not be deleted from the reception queue 11B. It is shown at (5) in
At (6) the packet receiver 15B receives data modification F1. The packet receiver 15B examines the values of the BT and the SBT included in the data modification F1 in order to add the data modification F1 after the data modification O4 in the reception queue 11B. The replica controller 7B determines whether the data modification F1 can be executed. Since the value of the SBT in the data modification F1 does not match the value of the SBT in process, the data modification F1 can not be executed, and accordingly, can not be deleted from the reception queue 11B. (6) in
At (7) the packet receiver 15B receives a data modification O3. The packet receiver 15B examines the values of the BT and the SBT included in the data modification O3 and rearranges the order in the reception queue 11B. As shown at (7) in
Furthermore, since the values of the BT and the SBT in the data modification O4 match respectively the values of the BT in process and the SBT in process, and since Ordinary is designated for the updating order in the data modification O3, the replica controller 7B increments by one the value of the SBT in process. As a result, the values of the BT and the SBT included in the data modification F1 that could not be executed at (6) match the values of the BT in process and the SBT in process, respectively. The replica controller 7B executes the data modification F1, and deletes it from the reception queue 11B. The rightmost state at (7) in
The overall processing described above is shown in FIG. 6. First, if a modification request is received and added to the reception queue 11 or if an instruction is received from step 2200, it is checked whether the modification request designated at step 2000 can be executed (step 2100). This determination is made by the designation of the updating order and the values of the BT and the SBT, all of which are included in the modification request. If the modification request can not be executed, processing of the succeeding modification request stored in the reception queue 11 is begun (step 2200). When there is no succeeding modification request in the reception queue 11, program control waits for the reception of a modification request. Although not shown in
As described above, the packet receiver 15 updates the order of the modification requests in the reception queue 11 each time such a request is received, and decides the timing/order of the execution of the individual modification requests by determining whether the modification requests can be executed. This operation is not at all related to the reception order. Even if the modification requests are not received in the generation order of the requests, an executable modification request can be identified and executed, so that the processing efficiency is improved. In addition, since some modification requests can be deleted after being executed, the efficiency of the use of the memory can be improved. Even if the executed modification request can not be deleted, only the basic information for the request need be maintained, and the efficiency in the use of memory is improved even more.
Only two computers are provided for the data sharing system in
The above described embodiment can be variously modified, and the components need not be equipped on the same computer. The functions of the packet receiver 15 and the replica controller 7 can be assigned to either of the components, as needed, and the components may be integrally formed. Similarly, the data storage device 17 and the replica controller 7 may be integrally formed. In the above embodiment, the execution and the deletion of a modification request are independently performed, but these may be performed at the same time. At this time, the values of the BT and the SBT have to be stored separately in order to update the values of the response BT and the response SBT. In addition, instead of managing the values of the response confirmation BT and the response confirmation SBT, each time the values of the response BT and the response SBT are received, a modification request that includes the equivalent values of the BT and the SBT may be deleted. The transmission queue and the reception queue may be stored in the main memory or in a dedicated memory. The modification request that has been executed may not be held in the reception queue, and another queue or table may be used to update the values of the BT and the SBT in process. Further, instead of managing the value of the SBT in process, modification requests in the reception queue including the same value of BT may be counted as needed.
The invention provides the following advantages:
1) A method and an apparatus for improving communication efficiency during the updating of a replica can be provided.
2) A method and an apparatus for reducing the amount of memory required for the updating of a replica can be provided.
3) A method and an apparatus for determining to which replica updating by a modification request is performed, even if a communication line is disconnected before updating the replica is completed can be provided.
4) The order in which a replica can be updated is clearly designated.
For example, a plurality of sheets of large image data are replicated, and the image data is divided into a plurality of segments to transmit the image data. In this case, the transmission order of the image data segments need not be controlled. However, it is preferable that, even if the communication line is disconnected intentionally or by accident during the replicating, image data is obtained for which as many data segments as possible are replicated. If the conventional FIFO order is used to replicate the image data, it can be confirmed that the image data have been replicated so long as data at the end of each image data segment was replicated. However, in this case the transmission overhead is large. In addition, even if the image data are replicated in a batch processing, once a line is disconnected, it can not be ensured that all the data in each image data segment have been replicated. According to the present invention, if Forward Flush is designated at the end of one image data set, the transmission order is not restricted as to the data transmission in one image data, and the data transmission order for images can be controlled. If the end data is updated, it is ensured that overall the image data have been replicated.
An explanation will be given for an example where personal information owned by a salesperson and his or her monthly sales chart are to be replicated. Assume that the personal information includes the employee's number and name, department and section, and telephone number, and that the monthly sales chart includes the employee's number and daily sales. A sales record output program obtains the employee's number from the monthly sales chart in order to search for the personal information, and outputs the corresponding name, department and section, and telephone number. In this example, if the personal information does not exist even though the corresponding monthly sales chart is present, the output program can not output the name, etc., that corresponds to the monthly sales chart. When a plurality of data sets that depend on each other are to be replicated, it is sometimes desirable that the order of the replicating data that depend on each other is ensured, so that no contradiction occurs in the dependency even when a communication line is disconnected during the replicating. When the conventional FIFO order is used for replicating, the replicating order is ensured by the generation of the personal information and the data of the monthly sales chart in the named order. However, other data of which the transmission order need not be ensured are also replicated in the FIFO order, and an unnecessary transmission overhead may be assessed. When the batch processing is performed for replicating, the replicating order can not be obtained if the communication line is disconnected before the replicating has been completed. According to the present invention, Two Way Flush is designated for the generation of the final personal information, and then the monthly sales chart is prepared. As a result, the order in which the personal information and the monthly sales chart are replicated is ensured. The transmission overhead that assessed during this process can be reduced to the minimum possible.
Negishi, Yasushi, Tago, Kazuya, Ahuja, Mohan
Patent | Priority | Assignee | Title |
6859821, | Jul 19 1999 | Microsoft Technology Licensing, LLC | Method and apparatus for prioritizing data change requests and maintaining data consistency in a distributed computer system equipped for activity-based collaboration |
7200869, | Sep 15 2000 | Microsoft Technology Licensing, LLC | System and method for protecting domain data against unauthorized modification |
7328273, | Mar 26 2001 | DUAXES Corporation | Protocol duplicating configuration and method for replicating and order of reception of packets on a first and second computers using a shared memory queue |
7490126, | Jul 19 1999 | Microsoft Technology Licensing, LLC | Method and apparatus for generating data change requests containing data consistency information in a peer-to-peer collaborative computer system |
7516204, | Feb 28 2003 | Canon Kabushiki Kaisha | Information processing method and apparatus |
7581069, | Oct 25 2005 | Waratek Limited | Multiple computer system with enhanced memory clean up |
7584243, | Nov 29 2002 | Canon Kabushiki Kaisha Kaisha | Information processing method and apparatus maintaining consistency of shared data |
7660960, | Oct 25 2005 | Waratek Limited | Modified machine architecture with partial memory updating |
7698521, | Jan 03 2002 | GOOGLE LLC | Data synchronization of multiple remote storage |
7707179, | Apr 23 2004 | Waratek Limited | Multiple computer architecture with synchronization |
7739240, | Dec 09 2002 | Hewlett Packard Enterprise Development LP | Replication and replica management in a wide area file system |
7757114, | Jan 03 2002 | Hitachi, Ltd. | Data synchronization of multiple remote storage after remote copy suspension |
7761670, | Oct 25 2005 | Waratek Limited | Modified machine architecture with advanced synchronization |
7788314, | Apr 20 2004 | Waratek Limited | Multi-computer distributed processing with replicated local memory exclusive read and write and network value update propagation |
7818296, | Apr 23 2004 | Waratek Limited | Computer architecture and method of operation for multi-computer distributed processing with synchronization |
7831779, | Oct 05 2006 | Waratek Limited | Advanced contention detection |
7844665, | Apr 23 2004 | Waratek Limited | Modified computer architecture having coordinated deletion of corresponding replicated memory locations among plural computers |
7849151, | Oct 05 2006 | Waratek Pty Ltd. | Contention detection |
7849369, | Oct 25 2005 | Waratek Limited | Failure resistant multiple computer system and method |
7849452, | Apr 23 2004 | Waratek Limited | Modification of computer applications at load time for distributed execution |
7852845, | Oct 05 2006 | Waratek Pty Ltd. | Asynchronous data transmission |
7860829, | Apr 23 2004 | Waratek Limited | Computer architecture and method of operation for multi-computer distributed processing with replicated memory |
7894341, | Oct 05 2006 | Waratek Limited | Switch protocol for network communications |
7949837, | Oct 09 2006 | Waratek Limited | Contention detection and resolution |
7958322, | Oct 25 2005 | Waratek Limited | Multiple machine architecture with overhead reduction |
7958329, | Oct 09 2006 | Waratek Pty Ltd | Hybrid replicated shared memory |
7962697, | Oct 05 2006 | Waratek Limited | Contention detection |
7971005, | Oct 05 2006 | Waratek Limited | Advanced contention detection |
7996627, | Oct 25 2005 | Waratek Limited | Replication of object graphs |
8015236, | Oct 25 2005 | Waratek Limited | Replication of objects having non-primitive fields, especially addresses |
8024404, | Jul 19 1999 | Microsoft Technology Licensing, LLC | Method and apparatus for designating endpoints in a collaborative computer system to facilitate maintaining data consistency |
8028299, | Apr 23 2004 | Waratek Limited | Computer architecture and method of operation for multi-computer distributed processing with finalization of objects |
8086805, | Oct 05 2006 | Waratek Limited | Advanced contention detection |
8090926, | Oct 05 2006 | Waratek Limited | Hybrid replicated shared memory |
8095616, | Oct 05 2006 | Waratek Limited | Contention detection |
8122198, | Oct 25 2005 | Waratek Limited | Modified machine architecture with partial memory updating |
8122200, | Oct 25 2005 | Waratek Limited | Modified machine architecture with advanced synchronization |
8156075, | Dec 19 2002 | SYNCHRONOSS SOFTWARE IRELAND LIMITED | Method of automatically replicating data objects between a mobile device and a server |
8209393, | Oct 25 2005 | Waratek Limited | Multiple machine architecture with overhead reduction |
8316190, | Apr 23 2004 | Waratek Limited | Computer architecture and method of operation for multi-computer distributed processing having redundant array of independent systems with replicated memory and code striping |
8374966, | Aug 01 2002 | Oracle International Corporation | In memory streaming with disk backup and recovery of messages captured from a database redo stream |
8473564, | Oct 05 2006 | Waratek Limited | Contention detection and resolution |
9197486, | Aug 29 2008 | GOOGLE LLC | Adaptive accelerated application startup |
9654542, | Aug 29 2008 | GOOGLE LLC | Adaptive accelerated application startup |
Patent | Priority | Assignee | Title |
5434994, | May 23 1994 | International Business Machines Corporation | System and method for maintaining replicated data coherency in a data processing system |
5765171, | Dec 29 1995 | THE CHASE MANHATTAN BANK, AS COLLATERAL AGENT | Maintaining consistency of database replicas |
5787262, | Jun 26 1996 | Microsoft Technology Licensing, LLC | System and method for distributed conflict resolution between data objects replicated across a computer network |
5819020, | Oct 16 1995 | CARBONITE, INC | Real time backup system |
5893116, | Sep 30 1996 | Oracle International Corporation | Accessing network resources using network resource replicator and captured login script for use when the computer is disconnected from the network |
6049809, | Oct 30 1996 | Microsoft Technology Licensing, LLC | Replication optimization system and method |
6052718, | Jan 07 1997 | Cisco Technology, Inc | Replica routing |
EP794646, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Oct 22 1999 | International Business Machines Corporation | (assignment on the face of the patent) | / | |||
Nov 02 1999 | NEGISHI, YASUSHI | International Business Machines Corp | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 010537 | /0932 | |
Dec 13 1999 | AHUJA, MOHAN | International Business Machines Corp | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 010537 | /0932 | |
Dec 13 1999 | TAGO, KAZUYA | International Business Machines Corp | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 010537 | /0932 |
Date | Maintenance Fee Events |
Sep 26 2006 | M1551: Payment of Maintenance Fee, 4th Year, Large Entity. |
Jan 03 2011 | REM: Maintenance Fee Reminder Mailed. |
May 27 2011 | EXP: Patent Expired for Failure to Pay Maintenance Fees. |
Date | Maintenance Schedule |
May 27 2006 | 4 years fee payment window open |
Nov 27 2006 | 6 months grace period start (w surcharge) |
May 27 2007 | patent expiry (for year 4) |
May 27 2009 | 2 years to revive unintentionally abandoned end. (for year 4) |
May 27 2010 | 8 years fee payment window open |
Nov 27 2010 | 6 months grace period start (w surcharge) |
May 27 2011 | patent expiry (for year 8) |
May 27 2013 | 2 years to revive unintentionally abandoned end. (for year 8) |
May 27 2014 | 12 years fee payment window open |
Nov 27 2014 | 6 months grace period start (w surcharge) |
May 27 2015 | patent expiry (for year 12) |
May 27 2017 | 2 years to revive unintentionally abandoned end. (for year 12) |